I just launched the Corn Maze app for the Crow River Winery for 2024 (year 3).
There’s some more background about it in a previous post. But I had a bit more time to put into it this year then I have had in the past so I thought I’d share a bit about the improvements I’ve managed to make this year and some general thoughts on how the project is going.
How to Play Video
I realized I needed a video to explain to players how to play since I watched actual people try to figure this out. Even though these games are very simple, the mechanics of playing are not something people (my users) are used to.
Here’s my first version of the “How to play” video.
I’ll try to make a better video next year. I’m still not sure the best way to capture using an app on a cell phone in a video.
Roll Playing Game Layout
I probably spent most of my time this year trying to make the UX for the Labyrinth game better. I had a few goals:
- It should feel like playing an RPG
- Mobile first layout (touch instead of a mouse)
- Layout works in landscape or portrait orientation
- Light/Dark mode
- Stick to web standard form elements
Here’s a video showing the layout change for landscape or portrait orientation.
I used a CSS Framework called Pico. It’s great. It got me the light and dark mode for (mostly) free. It’s very light weight and did everything I expected it to.
I listened to a podcast where the guests talked about complicated CSS is and how ill suited it is for many tasks (like making a game’s UI). I’m not sure I would have agreed with him, but after I spent a lot of time trying to make this UI I’m a lot more sympathetic. There has to be a better way, but maybe there isn’t. If you’re going to handle the whole zoo of mobile screen sizes maybe you’re just not going to have a good time.
A New Quiz (The LLM Helps Out)
Last year I did a lot of work to make it easy to add additional quizzes. Alas I ran out of time before I could add another quiz. This year I finally did add another quiz. I wanted to do something thematic so I went with “Wine in the Movies”. To help me write it I started with ChatGTP. It kinda worked. I needed 15 questions. After a lot of back and forth I had about 7 or 8 questions that were pretty good. When I tried to get it to write more questions it went off the rails. The questions where either complete hallucinations (i.e. imagining things that weren’t actually in the movies) or were just bad quiz questions.
I tried some other chat bots and didn’t have any better luck. After a lot of time trying to get them to be helpful I eventually gave up and just wrote the last questions I needed my self. Writing quiz questions is harder than I realized.
Overall it was a good learning experience. I’ve heard from others who use LLMs more than I do that, to make the best use of LLMs, know when to quit (and just do it your self). Give it 2 or 3 chances. If it’s not getting it, then just give up and do it your self.
But I like my framework for the quiz. I will try to add a new quiz every year.
Some Early Feedback
After some early feedback, it’s still too hard to figure out what to do. I’d chalk that up to it just being not a normal video game. I added some additional help but perhaps in the future I need to have some sort of a tutorial.
QR Codes
Sometimes the QR codes are not being read very quickly. You have to stand there and move the phone around to try to get it to read the QR codes. I wonder if I should try to find a better QR code library, or make them a different size. I wish there was an alternative that was less analog, like a bluetooth beacon, or an NFC tag. As far as I can tell though, there’s no good support for reading those inside of web apps.
2025 and Beyond
I reread my previous blog post on this. It’s interesting to reflect on some of the areas where I’ve made progress and the areas where I haven’t really.
User Experience
I do like the idea of better tools to make content, but I should probably spend a bit more time focusing on the UX. Maybe I can find a web framework that will make game UX easier/better.
Sound
I think for the QR code scanning and in a few other places in the app, having some sound feedback could be useful. I’m not sure how well that’s going to work in a web app, but I am to try to find out.